Comparing the performance of improved classify-analyze approaches for distal outcomes in latent profile analysis

John Dziak, Bethany Cara Bray, Jieting Zhang, Minqiang Zhang, Stephanie Trea Lanza

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Several approaches are available for estimating the relationship of latent class membership to distal outcomes in latent profile analysis (LPA). A three-step approach is commonly used, but has problems with estimation bias and confidence interval coverage. Proposed improvements include the correction method of Bolck, Croon, and Hagenaars (BCH; 2004), Vermunt's (2010) maximum likelihood (ML) approach, and the inclusive three-step approach of Bray, Lanza, and Tan (2015). These methods have been studied in the related case of latent class analysis (LCA) with categorical indicators, but not as well studied for LPA with continuous indicators. We investigated the performance of these approaches in LPA with normally distributed indicators, under different conditions of distal outcome distribution, class measurement quality, relative latent class size, and strength of association between latent class and the distal outcome. The modified BCH implemented in Latent GOLD had excellent performance. The maximum likelihood and inclusive approaches were not robust to violations of distributional assumptions. These findings broadly agree with and extend the results presented by Bakk and Vermunt (2016) in the context of LCA with categorical indicators.

Original languageEnglish (US)
Pages (from-to)107-116
Number of pages10
JournalMethodology
Volume12
Issue number4
DOIs
StatePublished - Oct 1 2016

Fingerprint

performance
Confidence Intervals
class membership
confidence
coverage
trend

All Science Journal Classification (ASJC) codes

  • Social Sciences(all)
  • Psychology(all)

Cite this

@article{6366b48e8e1047c5af1f003e7dc680f3,
title = "Comparing the performance of improved classify-analyze approaches for distal outcomes in latent profile analysis",
abstract = "Several approaches are available for estimating the relationship of latent class membership to distal outcomes in latent profile analysis (LPA). A three-step approach is commonly used, but has problems with estimation bias and confidence interval coverage. Proposed improvements include the correction method of Bolck, Croon, and Hagenaars (BCH; 2004), Vermunt's (2010) maximum likelihood (ML) approach, and the inclusive three-step approach of Bray, Lanza, and Tan (2015). These methods have been studied in the related case of latent class analysis (LCA) with categorical indicators, but not as well studied for LPA with continuous indicators. We investigated the performance of these approaches in LPA with normally distributed indicators, under different conditions of distal outcome distribution, class measurement quality, relative latent class size, and strength of association between latent class and the distal outcome. The modified BCH implemented in Latent GOLD had excellent performance. The maximum likelihood and inclusive approaches were not robust to violations of distributional assumptions. These findings broadly agree with and extend the results presented by Bakk and Vermunt (2016) in the context of LCA with categorical indicators.",
author = "John Dziak and Bray, {Bethany Cara} and Jieting Zhang and Minqiang Zhang and Lanza, {Stephanie Trea}",
year = "2016",
month = "10",
day = "1",
doi = "10.1027/1614-2241/a000114",
language = "English (US)",
volume = "12",
pages = "107--116",
journal = "Methodology",
issn = "1614-1881",
publisher = "Hogrefe Publishing",
number = "4",

}

Comparing the performance of improved classify-analyze approaches for distal outcomes in latent profile analysis. / Dziak, John; Bray, Bethany Cara; Zhang, Jieting; Zhang, Minqiang; Lanza, Stephanie Trea.

In: Methodology, Vol. 12, No. 4, 01.10.2016, p. 107-116.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Comparing the performance of improved classify-analyze approaches for distal outcomes in latent profile analysis

AU - Dziak, John

AU - Bray, Bethany Cara

AU - Zhang, Jieting

AU - Zhang, Minqiang

AU - Lanza, Stephanie Trea

PY - 2016/10/1

Y1 - 2016/10/1

N2 - Several approaches are available for estimating the relationship of latent class membership to distal outcomes in latent profile analysis (LPA). A three-step approach is commonly used, but has problems with estimation bias and confidence interval coverage. Proposed improvements include the correction method of Bolck, Croon, and Hagenaars (BCH; 2004), Vermunt's (2010) maximum likelihood (ML) approach, and the inclusive three-step approach of Bray, Lanza, and Tan (2015). These methods have been studied in the related case of latent class analysis (LCA) with categorical indicators, but not as well studied for LPA with continuous indicators. We investigated the performance of these approaches in LPA with normally distributed indicators, under different conditions of distal outcome distribution, class measurement quality, relative latent class size, and strength of association between latent class and the distal outcome. The modified BCH implemented in Latent GOLD had excellent performance. The maximum likelihood and inclusive approaches were not robust to violations of distributional assumptions. These findings broadly agree with and extend the results presented by Bakk and Vermunt (2016) in the context of LCA with categorical indicators.

AB - Several approaches are available for estimating the relationship of latent class membership to distal outcomes in latent profile analysis (LPA). A three-step approach is commonly used, but has problems with estimation bias and confidence interval coverage. Proposed improvements include the correction method of Bolck, Croon, and Hagenaars (BCH; 2004), Vermunt's (2010) maximum likelihood (ML) approach, and the inclusive three-step approach of Bray, Lanza, and Tan (2015). These methods have been studied in the related case of latent class analysis (LCA) with categorical indicators, but not as well studied for LPA with continuous indicators. We investigated the performance of these approaches in LPA with normally distributed indicators, under different conditions of distal outcome distribution, class measurement quality, relative latent class size, and strength of association between latent class and the distal outcome. The modified BCH implemented in Latent GOLD had excellent performance. The maximum likelihood and inclusive approaches were not robust to violations of distributional assumptions. These findings broadly agree with and extend the results presented by Bakk and Vermunt (2016) in the context of LCA with categorical indicators.

UR - http://www.scopus.com/inward/record.url?scp=85002913948&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85002913948&partnerID=8YFLogxK

U2 - 10.1027/1614-2241/a000114

DO - 10.1027/1614-2241/a000114

M3 - Article

VL - 12

SP - 107

EP - 116

JO - Methodology

JF - Methodology

SN - 1614-1881

IS - 4

ER -